AI Answer Writing & Review

I’m building a dataset of short, question-and-answer pairs that will be used to train and fine-tune large language models. I need several entry-level assistants who can (1) draft concise answers to a variety of everyday questions and (2) review AI-generated drafts for accuracy and readability.

Here’s how the workflow looks: I’ll give you batches of prompts. For each prompt you’ll either create a fresh answer or polish an answer the model already produced. Your focus is to keep the facts correct, wording clear, and tone friendly yet natural—no stiff “robot speak.” If something feels off, flag it and suggest an improved version.

Typical guidelines
• 1–3 sentences per answer unless the prompt demands more.
• Check names, dates, and figures against reliable open sources.
• Rewrite any awkward phrasing so it sounds human and helpful.
• Maintain neutrality—no personal opinions unless the question calls for it.

Tooling
Work happens in a shared Google Sheet and, for longer batches, a simple web interface similar to OpenAI Playground. Familiarity with Grammarly or similar checkers is a plus, but not mandatory.

Deliverables
– Completed batch of Q&A entries returned by the agreed deadline
– Change log or comments on any corrections you make
– Brief note summarizing common issues you observed (helps refine the next prompt set)

Acceptance criteria
An entry passes when it is factually correct, free of grammar errors, and reads as if a helpful human wrote it. Random spot checks will be run with plagiarism scanners and manual review.

If you’re detail-oriented, comfortable fact-checking quickly, and curious about AI training, this should be straightforward and steady work. Let me know your availability and how many answers per day you can handle.